ralphex
Open-source desktop app for Mac, Windows and Linux. Agents run on your computer or on your own server.
Product websitegithub.com Compare with others Missing or wrong data?
- Desktop app
- Mac, Windows and Linux
- Browser
- Yes
- Phone
- Phone browser
- Command line
- Yes
- Agents
- Claude Code, Codex, Cursor Agent, Gemini CLI, OpenCode, Pi and Antigravity
- Your subscription
- Works with your own plan
- Agents run
- On your computer and on your own server
- Laptop closed
- Keeps going
- Isolation
- Worktree, Container and VM
- Parallel agents
- Yes
- Scheduled runs
- No
- Coordinator agent
- No
- Built-in issues
- Yes
- Pull requests
- No
- Review screen
- No
- Team
- No
- Self-hosting
- Yes
- Licence
- MIT
- Price from
- Free
- How it charges
- No paid plan and bring your own key
- Paid plans from
- Not listed
Receipts
- Checked
- · oldest answer 2026-11-08
- Findings
- 37 published · 25 cross-checked by a second agent · read them all
- Sources
- 16 public pages on 2 sites: github.com, ralphex.com
- Screenshot
- github.com, retrieved · first party repository · shows appearance, not capability
- Data
- /products/ralphex.json · .md
- Method
- Same rules for all 103 researched ADEs. How we check · Report wrong data
What it does
Where you use it
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
Source and scope for Where you use itAgents it runs
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
Source and scope for Agents it runsAccounts and subscriptions
Provider authentication is inherited from agent CLIs; Claude headless subscription use has a separate billing warning.
Source and scope for Accounts and subscriptionsWhere agents run
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
Source and scope for Where agents runHow work is kept apart
Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
Source and scope for How work is kept apartParallel work
Parallel review agents and independent worktree plan runs are documented.
Source and scope for Parallel workScheduled runs
The inspected Ralphex runtime starts execution through a human CLI invocation. Its server watches progress and outbound notifications report completion; neither supplies unattended run triggers.
Source and scope for Scheduled runsGit and pull requests
Ralphex creates branches/worktrees and commits, and its optional finalize prompt rebases/organizes commits. It has no native PR creation or management operation.
Source and scope for Git and pull requestsPrice
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
Source and scope for PriceTyped filters
Every value links to its receipt below. Read plan and surface restrictions before using a filter to decide.
| Filter | Value |
|---|---|
| Desktop OS: macOS | Yes |
| Desktop OS: Windows | Yes |
| Desktop OS: Linux | Yes |
| Web client | Yes |
| Phone access: Native iOS | No |
| Phone access: Native Android | No |
| Phone access: Phone browser | Yes |
| CLI | Yes cross-checked |
| Coding agents: Claude Code | Yes |
| Coding agents: Codex | Yes |
| Coding agents: Cursor Agent | Yes |
| Coding agents: Gemini CLI | Yes |
| Coding agents: OpenCode | Yes |
| Coding agents: Pi | Yes |
| Coding agents: Grok | Unknown |
| Coding agents: Aider | Unknown |
| Coding agents: Antigravity | Yes |
| Use my subscription | Yes cross-checked |
| Where agents run: On your computer | Yes |
| Where agents run: On your own server | Yes |
| Where agents run: In the vendor's cloud | No |
| Workspace isolation: Git worktree | Yes |
| Workspace isolation: Container | Yes |
| Workspace isolation: VM or microVM | Yes |
| Parallel runs | Yes cross-checked |
| Coordinator | No |
| Native issue tracking | Yes |
| Unattended triggers | No |
| Shared prompting | No |
| Team roles | No |
| SSO | No |
| Pull request workflow | No |
| Review interface | No |
| Product license: Apache-2.0 | No. MIT cross-checked |
| Product license: MIT | Yes. MIT cross-checked |
| Product license: GPL-3.0 | No. MIT cross-checked |
| Product license: GPL-3.0-or-later | No. MIT cross-checked |
| Product license: AGPL-3.0 | No. MIT cross-checked |
| Product license: AGPL-3.0-or-later | No. MIT cross-checked |
| Product license: Elastic License 2.0 | No. MIT cross-checked |
| Product license: Custom terms | No. MIT cross-checked |
| Product license: Proprietary | No. MIT cross-checked |
| Self-hosting | Yes cross-checked |
| Telemetry opt-out | No |
| Free access | Yes cross-checked |
| Starting price: Free | Yes. Free |
| Starting price: Up to $20/month | No. Free |
| Starting price: $21–40/month | No. Free |
| Starting price: $41–100/month | No. Free |
| Starting price: Over $100/month | No. Free |
| Starting price: Contact for pricing | No. Free |
| Pricing model: Free, no paid plan | Yes cross-checked |
| Pricing model: Free plan | No |
| Pricing model: Subscription | No |
| Pricing model: Per seat | No |
| Pricing model: Usage-based | No |
| Pricing model: Per machine or compute | No |
| Pricing model: Bring your own key | Yes cross-checked |
| Pricing model: One-time purchase | No |
| Paid plans from: No paid plan | Unknown |
| Paid plans from: $10 or less/month | Unknown |
| Paid plans from: $11–20/month | Unknown |
| Paid plans from: $21–40/month | Unknown |
| Paid plans from: $41–100/month | Unknown |
| Paid plans from: Over $100/month | Unknown |
| Paid plans from: Contact sales | Unknown |
| Lifecycle status: Active | Unknown |
| Lifecycle status: Sunsetting | Unknown |
| Keeps going with laptop closed | Yes |
| Open-source license | Yes. MIT cross-checked |
| Any desktop app | Yes |
| Paid plan up to $20/month | Unknown |
| Paid plan up to $40/month | Unknown |
| Any phone client | Yes |
Filter receipts
Desktop OS: macOS: Yes
Desktop OS: macOS
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Desktop OS: Windows: Yes
Desktop OS: Windows
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Desktop OS: Linux: Yes
Desktop OS: Linux
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Web client: Yes
Web client
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Desktop os: Linux/macOS supported; Windows best-effort source-only, no release binary or maintainer tests.
- Browser: Optional --serve dashboard monitors progress; steering via browser not established.
- Cli: Standalone ralphex binary.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
Phone access: Native iOS: No
Phone access: Native iOS
Typed value: No.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Phone access: Native Android: No
Phone access: Native Android
Typed value: No.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Phone access: Phone browser: Yes
Phone access: Phone browser
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
CLI: Yes cross-checked
CLI
Typed value: Yes.
CLI configures plan creation, execution, review, worktrees and dashboard.
- Cli: --plan, --tasks-only, --review, --external-only, --codex, --worktree, --serve.
- Operations: Start execution and configure phase behavior/limits.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Coding agents: Claude Code: Yes
Coding agents: Claude Code
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Custom adapters: claude_command/claude_args allow a compatible stream-JSON wrapper.
- Session integration: Each implementation task uses a fresh coding-agent session.
- Mixed harness operation: Default pipeline uses Claude for implementation and optional Codex for external review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
Coding agents: Codex: Yes
Coding agents: Codex
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Custom adapters: claude_command/claude_args allow a compatible stream-JSON wrapper.
- Session integration: Each implementation task uses a fresh coding-agent session.
- Mixed harness operation: Default pipeline uses Claude for implementation and optional Codex for external review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
Coding agents: Cursor Agent: Yes
Coding agents: Cursor Agent
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
Finding: The first-party README explicitly documents Cursor CLI as a community-tested drop-in agent: claude_command=agent and claude_args=--force --output-format stream-json. Ralphex launches it through the task/review executor. Compatibility depends on Cursor retaining the expected stream format and is not officially supported.
Coding agents: Gemini CLI: Yes
Coding agents: Gemini CLI
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Custom adapters: claude_command/claude_args allow a compatible stream-JSON wrapper.
- Session integration: Each implementation task uses a fresh coding-agent session.
- Mixed harness operation: Default pipeline uses Claude for implementation and optional Codex for external review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
Coding agents: OpenCode: Yes
Coding agents: OpenCode
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Custom adapters: claude_command/claude_args allow a compatible stream-JSON wrapper.
- Session integration: Each implementation task uses a fresh coding-agent session.
- Mixed harness operation: Default pipeline uses Claude for implementation and optional Codex for external review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
Coding agents: Pi: Yes
Coding agents: Pi
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Custom adapters: claude_command/claude_args allow a compatible stream-JSON wrapper.
- Session integration: Each implementation task uses a fresh coding-agent session.
- Mixed harness operation: Default pipeline uses Claude for implementation and optional Codex for external review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
Coding agents: Antigravity: Yes
Coding agents: Antigravity
Typed value: Yes.
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Harnesses: Claude Code/Codex and the documented named wrappers include Antigravity; Cursor CLI is explicitly configured as the launched task/review command, with community-tested compatibility.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
ralphex Antigravity support at v1.7.0
Finding: Ralphex v1.7.0 ships agy-as-claude.sh, a configured replacement for Claude Code in task and review phases that translates Antigravity output into Claude stream-json.
Use my subscription: Yes cross-checked
Use my subscription
Typed value: Yes.
Provider authentication is inherited from agent CLIs; Claude headless subscription use has a separate billing warning.
- Agent subscriptions: README explicitly addresses Claude subscription users and separate Agent SDK credit pool.
- Api accounts: README documents API-key, Bedrock, Vertex and Foundry routes.
- Provider restrictions: Default uses claude --print; unofficial interactive wrappers are not guaranteed stable.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Where agents run: On your computer: Yes
Where agents run: On your computer
Typed value: Yes.
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
- Local: Local CLI with no required ralphex cloud service.
- Payer: User provides agent accounts and host compute.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, resources, region, expiry, other_device_continuation.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Where agents run: On your own server: Yes
Where agents run: On your own server
Typed value: Yes.
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
- Byo remote: The CLI exposes --serve --host/RALPHEX_WEB_HOST and documented Docker execution for an operator-controlled host.
- Managed cloud: The complete inspected execution modes run on operator-owned hosts; no Ralphex-operated execution service is provided.
- Client disconnect: The browser only observes SSE output; agent lifetime belongs to the installed CLI host.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, resources, region, expiry, other_device_continuation.
-
Finding: The CLI deliberately serves a network monitoring dashboard using --serve --host or RALPHEX_WEB_HOST. The installed host owns agent execution and the browser is an SSE observer, so a separately deployed operator host/Docker host can continue work with the client laptop disconnected. No Ralphex-operated cloud runtime is implemented in the inspected execution paths.
Where agents run: In the vendor's cloud: No
Where agents run: In the vendor's cloud
Typed value: No.
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
- Byo remote: The CLI exposes --serve --host/RALPHEX_WEB_HOST and documented Docker execution for an operator-controlled host.
- Managed cloud: The complete inspected execution modes run on operator-owned hosts; no Ralphex-operated execution service is provided.
- Client disconnect: The browser only observes SSE output; agent lifetime belongs to the installed CLI host.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, resources, region, expiry, other_device_continuation.
-
Finding: The CLI deliberately serves a network monitoring dashboard using --serve --host or RALPHEX_WEB_HOST. The installed host owns agent execution and the browser is an SSE observer, so a separately deployed operator host/Docker host can continue work with the client laptop disconnected. No Ralphex-operated cloud runtime is implemented in the inspected execution paths.
Workspace isolation: Git worktree: Yes
Workspace isolation: Git worktree
Typed value: Yes.
Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
- Branch: Feature branch created from plan filename.
- Worktree: --worktree uses .ralphex/worktrees/<branch>.
- Container: Docker wrapper runs the CLI and agent processes in a container.
- Secrets: Docker credential mounts are documented; not a hostile-code security guarantee.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: checkout, os_sandbox, filesystem, process, network, code_separation, security_boundary.
-
Finding: Optional --worktree runs plans in separate Git worktrees; Docker wrapper mounts project/credentials with documented sandbox limits. Worktrees only apply to full/tasks-only modes.
Workspace isolation: Container: Yes
Workspace isolation: Container
Typed value: Yes.
Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
- Branch: Feature branch created from plan filename.
- Worktree: --worktree uses .ralphex/worktrees/<branch>.
- Container: Docker wrapper runs the CLI and agent processes in a container.
- Secrets: Docker credential mounts are documented; not a hostile-code security guarantee.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: checkout, os_sandbox, filesystem, process, network, code_separation, security_boundary.
-
Finding: Optional --worktree runs plans in separate Git worktrees; Docker wrapper mounts project/credentials with documented sandbox limits. Worktrees only apply to full/tasks-only modes.
Workspace isolation: VM or microVM: Yes
Workspace isolation: VM or microVM
Typed value: Yes.
Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
- Vm: Optional Docker-on-macOS runs inside the operator Docker VM; this is not task-specific Ralphex VM provisioning.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: checkout, os_sandbox, filesystem, process, network, code_separation, security_boundary.
-
Finding: The optional Docker wrapper documents that macOS Docker has VM isolation, in contrast to the Linux host-socket security warning. This is the operator Docker-on-macOS VM boundary, not a per-task Ralphex VM provisioning service.
Parallel runs: Yes cross-checked
Parallel runs
Typed value: Yes.
Parallel review agents and independent worktree plan runs are documented.
- Concurrent sessions: First review launches five provider-native reviewers in parallel.
- Concurrent tasks: Multiple plans can execute concurrently in separate --worktree runs.
- Status visibility: Streaming log and optional dashboard with phase filters.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Coordinator: No
Coordinator
Typed value: No.
Ralphex owns deterministic task/review phases and ships reviewer prompts, but delegation uses Claude Task or Codex spawn_agent rather than Ralphex-owned session delegation tools.
- Messages: No Ralphex-owned tool for a coordinator to message or delegate to other agent sessions exists in the inspected prompt/runtime surface.
- Orchestration owner: Ralphex supplies the review pipeline and prompts, while reviewer delegation is performed by provider-native tools.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_harness_effort, authentication, live_progress, retained_results, followup, invocation.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Native issue tracking: Yes
Native issue tracking
Typed value: Yes.
Ralphex owns persistent Markdown plan tasks and checkbox state, separate from each coding-agent execution session. Plans can be created, edited, executed later and archived on completion.
- Create: --plan creates a persistent plan file; manually authored/imported plans are also accepted.
- Edit: The plan review flow supports text feedback and $EDITOR annotation; the persistent Markdown can be edited.
- Archive: Successful completion moves the plan into the completed archive.
- Explicit dispatch: A CLI invocation selects a saved plan for execution.
- Task session separation: Users can exit after plan creation and execute it later; fresh agent sessions consume the file.
- Durable states: Task/Checkbox objects derive pending/active/done state from persisted plan checkboxes.
- External tracker requirement: The native task workflow uses local plan files rather than an external tracker.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: discuss, prioritize, label, assign, search, filter, reopen, project_epic_views, idea_parking, metadata.
-
Finding: Ralphex creates or imports Markdown plan files independently of an execution session. Task sections and persisted checkboxes are parsed into product task objects and pending/active/done states; users can edit the file, exit after planning, and execute/resume it later. Completed plans are archived. This is a native file-backed task workflow, not an external tracker link or chat-only list.
Unattended triggers: No
Unattended triggers
Typed value: No.
The inspected Ralphex runtime starts execution through a human CLI invocation. Its server watches progress and outbound notifications report completion; neither supplies unattended run triggers.
- Schedules: No inbound scheduled execution path exists in the inspected CLI/runtime.
- Events: The HTTP dashboard only reads progress/plan/session data; configured webhooks are outbound completion notifications.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: recurring_tasks, completion_conditions, scheduler_location, overlaps, retries, client_independence.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Shared prompting: No
Shared prompting
Typed value: No.
The inspected Ralphex interfaces are a single terminal operator and read-only browser observers; the dashboard cannot send human instructions into agent work.
- Coprompting: No shared-human prompt mutation path exists in the CLI/dashboard route surface.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: visibility, discussion, comments, presence, ordering, conflicts, attribution, permissions, plan_restrictions, owner_online.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Team roles: No
Team roles
Typed value: No.
The inspected Ralphex CLI/dashboard has no human membership, named human role permissions or enterprise SSO implementation.
- Organization roles: The source has no human workspace membership/role permission model.
- Sso: The complete local CLI/dashboard surface contains no SAML/OIDC SSO implementation.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: project_roles, task_roles, membership, provisioning, unauthorized_access, revocation, plan_scope.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
SSO: No
SSO
Typed value: No.
The inspected Ralphex CLI/dashboard has no human membership, named human role permissions or enterprise SSO implementation.
- Organization roles: The source has no human workspace membership/role permission model.
- Sso: The complete local CLI/dashboard surface contains no SAML/OIDC SSO implementation.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: project_roles, task_roles, membership, provisioning, unauthorized_access, revocation, plan_scope.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Pull request workflow: No
Pull request workflow
Typed value: No.
Ralphex creates branches/worktrees and commits, and its optional finalize prompt rebases/organizes commits. It has no native PR creation or management operation.
- Pull requests: The complete command/dashboard surface and shipped finalize prompt have no PR creation/management path.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: comments, ci_response, conflicts, product_state.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Review interface: No
Review interface
Typed value: No.
The Ralphex dashboard displays task progress, logs and aggregate change counts; it does not render proposed-code diffs for human review.
- Change understanding: diffStats renders changed-file/addition/deletion counts only, without code hunks or a proposed-change renderer.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: pending_reviews, evidence_navigation, comments, request_changes, shipping_guidance, effort, missed_evidence, keyboard, narrow_screen, context_switching, roles, retained_context.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Product license: Apache-2.0: No. MIT cross-checked
Product license: Apache-2.0
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: MIT: Yes. MIT cross-checked
Product license: MIT
Typed value: Yes. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: GPL-3.0: No. MIT cross-checked
Product license: GPL-3.0
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: GPL-3.0-or-later: No. MIT cross-checked
Product license: GPL-3.0-or-later
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: AGPL-3.0: No. MIT cross-checked
Product license: AGPL-3.0
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: AGPL-3.0-or-later: No. MIT cross-checked
Product license: AGPL-3.0-or-later
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: Elastic License 2.0: No. MIT cross-checked
Product license: Elastic License 2.0
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: Custom terms: No. MIT cross-checked
Product license: Custom terms
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Product license: Proprietary: No. MIT cross-checked
Product license: Proprietary
Typed value: No. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Self-hosting: Yes cross-checked
Self-hosting
Typed value: Yes.
Core standalone CLI requires no product cloud service.
- Self hosting: Run local binary or Docker on user-controlled infrastructure.
- Vendor dependencies: Agent/provider accounts required separately.
- Account independence: No IDE plugin or ralphex cloud service is required.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Telemetry opt-out: No
Telemetry opt-out
Typed value: No.
The inspected Ralphex configuration/runtime has no ADE analytics reporter or documented analytics opt-out control. Agent-provider traffic and optional user-configured completion notifications are separate.
- Telemetry disclosure: No ADE analytics reporter was found in the bounded runtime/configuration surface.
- Telemetry controls: The shipped configuration and CLI contain no ADE telemetry opt-out control.
- Provider separation: Selected agent/provider traffic and user-configured outbound notifications are separate from ADE analytics.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: code_data, prompts, credentials, logs, artifacts, retention, deletion, training, residency, subprocessors, telemetry_defaults, consent_prompt.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
Free access: Yes cross-checked
Free access
Typed value: Yes.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: Free: Yes. Free
Starting price: Free
Typed value: Yes. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: Up to $20/month: No. Free
Starting price: Up to $20/month
Typed value: No. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: $21–40/month: No. Free
Starting price: $21–40/month
Typed value: No. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: $41–100/month: No. Free
Starting price: $41–100/month
Typed value: No. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: Over $100/month: No. Free
Starting price: Over $100/month
Typed value: No. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Starting price: Contact for pricing: No. Free
Starting price: Contact for pricing
Typed value: No. Free.
Open-source binary and source allow evaluation with user-provided agents/compute.
- Ongoing free: MIT source and direct binary installation; no ralphex account required.
- Self host cost: Agent billing and compute remain the user-owned costs.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
Pricing model: Free, no paid plan: Yes cross-checked
Pricing model: Free, no paid plan
Typed value: Yes.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Free plan: No
Pricing model: Free plan
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Subscription: No
Pricing model: Subscription
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Per seat: No
Pricing model: Per seat
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Usage-based: No
Pricing model: Usage-based
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Per machine or compute: No
Pricing model: Per machine or compute
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Pricing model: Bring your own key: Yes cross-checked
Pricing model: Bring your own key
Typed value: Yes.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
Pricing model: One-time purchase: No
Pricing model: One-time purchase
Typed value: No.
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- Model cost: Model usage counts against your own Claude plan or Agent SDK credit pool, Anthropic API key billing, or OpenAI/codex plan.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
Keeps going with laptop closed: Yes
Keeps going with laptop closed
Typed value: Yes.
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
- Byo remote: The CLI exposes --serve --host/RALPHEX_WEB_HOST and documented Docker execution for an operator-controlled host.
- Managed cloud: The complete inspected execution modes run on operator-owned hosts; no Ralphex-operated execution service is provided.
- Client disconnect: The browser only observes SSE output; agent lifetime belongs to the installed CLI host.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, resources, region, expiry, other_device_continuation.
-
Finding: The CLI deliberately serves a network monitoring dashboard using --serve --host or RALPHEX_WEB_HOST. The installed host owns agent execution and the browser is an SSE observer, so a separately deployed operator host/Docker host can continue work with the client laptop disconnected. No Ralphex-operated cloud runtime is implemented in the inspected execution paths.
Open-source license: Yes. MIT cross-checked
Open-source license
Typed value: Yes. MIT.
Repository source is MIT licensed.
- Client license: MIT.
- Source visibility: Public repository.
- Source revision: dc16f39845b48d8f97a6884622e65d86d60e9040
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
Any desktop app: Yes
Any desktop app
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
Any phone client: Yes
Any phone client
Typed value: Yes.
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Mobile apps: The complete source client inventory is CLI and browser dashboard; there are no product-owned native iOS/Android clients.
- Mobile browser: The monitoring dashboard ships viewport metadata and specific small-screen layout rules.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
All public findings
Findings may summarize a source. They are not presented as verbatim statements.
ralphex:ACC-01: Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
ralphex:ACC-01
Ralphex is a primary native CLI on Linux/macOS and best-effort Windows source builds, with an optional responsive read-only browser monitoring dashboard and no native phone apps.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: editor_extensions, task_actions, notifications, cross_device_continuity.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
-
Finding: The primary client is the installed CLI. Linux/macOS are supported; Windows explicitly builds and runs from source on a best-effort basis, without release binaries or maintainer testing. The optional browser monitoring client has viewport metadata and deliberate 768px/640px phone layout rules; the source client inventory has no native iOS/Android application.
ralphex:ACC-02: Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
ralphex:ACC-02
Claude Code and Codex have first-class executors, several named agents have shipped wrappers, and the first-party README documents a community-tested Cursor CLI drop-in configuration.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_providers.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
-
ralphex Antigravity support at v1.7.0
Finding: Ralphex v1.7.0 ships agy-as-claude.sh, a configured replacement for Claude Code in task and review phases that translates Antigravity output into Claude stream-json.
-
Finding: The first-party README explicitly documents Cursor CLI as a community-tested drop-in agent: claude_command=agent and claude_args=--force --output-format stream-json. Ralphex launches it through the task/review executor. Compatibility depends on Cursor retaining the expected stream format and is not officially supported.
ralphex:ACC-03: Provider authentication is inherited from agent CLIs; Claude headless subscription use has a separate billing warning. cross-checked
ralphex:ACC-03
Provider authentication is inherited from agent CLIs; Claude headless subscription use has a separate billing warning.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:ACC-04: Run from a Git repository with a plan and installed coding agent. cross-checked
ralphex:ACC-04
Run from a Git repository with a plan and installed coding agent.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:ACC-05: CLI configures plan creation, execution, review, worktrees and dashboard. cross-checked
ralphex:ACC-05
CLI configures plan creation, execution, review, worktrees and dashboard.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:ACC-06: Custom prompts, review agents and provider configuration extend the workflow. cross-checked
ralphex:ACC-06
Custom prompts, review agents and provider configuration extend the workflow.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
ralphex:PLAN-01: Markdown plans are authoritative local execution input. cross-checked
ralphex:PLAN-01
Markdown plans are authoritative local execution input.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:PLAN-02: Plans retain validation commands, tasks and checkbox progress; paused work can be amended. cross-checked
ralphex:PLAN-02
Plans retain validation commands, tasks and checkbox progress; paused work can be amended.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:PLAN-03: Plan tasks execute sequentially with configurable multi-phase reviews. cross-checked
ralphex:PLAN-03
Plan tasks execute sequentially with configurable multi-phase reviews.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:PLAN-04: Repository guidance and review prompts can be configured per project. cross-checked
ralphex:PLAN-04
Repository guidance and review prompts can be configured per project.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
ralphex:PLAN-06: Ralphex owns persistent Markdown plan tasks and checkbox state, separate from each coding-agent execution session. Plans can be created, edited, executed later and archived on completion.
ralphex:PLAN-06
Ralphex owns persistent Markdown plan tasks and checkbox state, separate from each coding-agent execution session. Plans can be created, edited, executed later and archived on completion.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: discuss, prioritize, label, assign, search, filter, reopen, project_epic_views, idea_parking, metadata.
-
Finding: Ralphex creates or imports Markdown plan files independently of an execution session. Task sections and persisted checkboxes are parsed into product task objects and pending/active/done states; users can edit the file, exit after planning, and execute/resume it later. Completed plans are archived. This is a native file-backed task workflow, not an external tracker link or chat-only list.
ralphex:COORD-01: Parallel review agents and independent worktree plan runs are documented. cross-checked
ralphex:COORD-01
Parallel review agents and independent worktree plan runs are documented.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
ralphex:COORD-03: Ralphex owns deterministic task/review phases and ships reviewer prompts, but delegation uses Claude Task or Codex spawn_agent rather than Ralphex-owned session delegation tools.
ralphex:COORD-03
Ralphex owns deterministic task/review phases and ships reviewer prompts, but delegation uses Claude Task or Codex spawn_agent rather than Ralphex-owned session delegation tools.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: model_harness_effort, authentication, live_progress, retained_results, followup, invocation.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:COORD-04: The inspected Ralphex runtime starts execution through a human CLI invocation. Its server watches progress and outbound notifications report completion; neither supplies unattended run triggers.
ralphex:COORD-04
The inspected Ralphex runtime starts execution through a human CLI invocation. Its server watches progress and outbound notifications report completion; neither supplies unattended run triggers.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: recurring_tasks, completion_conditions, scheduler_location, overlaps, retries, client_independence.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:COORD-05: Unix signal controls pause/abort execution; notifications are optional. cross-checked
ralphex:COORD-05
Unix signal controls pause/abort execution; notifications are optional.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:COORD-06: Dashboard watches multiple plan progress files without a verified epic-management fixture. cross-checked
ralphex:COORD-06
Dashboard watches multiple plan progress files without a verified epic-management fixture.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
ralphex:COORD-07: Modes and per-phase model settings are configurable; workflow changes are mostly configuration-driven. cross-checked
ralphex:COORD-07
Modes and per-phase model settings are configurable; workflow changes are mostly configuration-driven.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
ralphex:RUN-01: Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
ralphex:RUN-01
Ralphex runs installed agent processes on the operator machine or Docker host. Its deliberately network-bindable monitoring server allows a separate browser laptop to disconnect while that host executes.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: self_hosted, hybrid, provisioning, resources, region, expiry, other_device_continuation.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: The CLI deliberately serves a network monitoring dashboard using --serve --host or RALPHEX_WEB_HOST. The installed host owns agent execution and the browser is an SSE observer, so a separately deployed operator host/Docker host can continue work with the client laptop disconnected. No Ralphex-operated cloud runtime is implemented in the inspected execution paths.
ralphex:RUN-02: Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
ralphex:RUN-02
Ralphex offers Git worktrees and a Docker wrapper. The Docker documentation explicitly identifies the macOS Docker VM boundary; Ralphex does not provision separate task VMs.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: checkout, os_sandbox, filesystem, process, network, code_separation, security_boundary.
-
Finding: Optional --worktree runs plans in separate Git worktrees; Docker wrapper mounts project/credentials with documented sandbox limits. Worktrees only apply to full/tasks-only modes.
-
Finding: The optional Docker wrapper documents that macOS Docker has VM isolation, in contrast to the Linux host-socket security warning. This is the operator Docker-on-macOS VM boundary, not a per-task Ralphex VM provisioning service.
ralphex:RUN-03: Plan progress and interrupted worktrees survive, but context is intentionally fresh per task. cross-checked
ralphex:RUN-03
Plan progress and interrupted worktrees survive, but context is intentionally fresh per task.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Optional --worktree runs plans in separate Git worktrees; Docker wrapper mounts project/credentials with documented sandbox limits. Worktrees only apply to full/tasks-only modes.
ralphex:RUN-05: Iteration, review-patience, timeout and rate-limit controls are documented. cross-checked
ralphex:RUN-05
Iteration, review-patience, timeout and rate-limit controls are documented.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:RUN-06: Timestamped logs and dashboard expose progress; this is not a tested state-consistency result. cross-checked
ralphex:RUN-06
Timestamped logs and dashboard expose progress; this is not a tested state-consistency result.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
ralphex:TEAM-01: The inspected Ralphex interfaces are a single terminal operator and read-only browser observers; the dashboard cannot send human instructions into agent work.
ralphex:TEAM-01
The inspected Ralphex interfaces are a single terminal operator and read-only browser observers; the dashboard cannot send human instructions into agent work.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: visibility, discussion, comments, presence, ordering, conflicts, attribution, permissions, plan_restrictions, owner_online.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:TEAM-02: The inspected Ralphex CLI/dashboard has no human membership, named human role permissions or enterprise SSO implementation.
ralphex:TEAM-02
The inspected Ralphex CLI/dashboard has no human membership, named human role permissions or enterprise SSO implementation.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: project_roles, task_roles, membership, provisioning, unauthorized_access, revocation, plan_scope.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:TEAM-03: Default Claude invocation bypasses permission prompts; safer host isolation is optional. cross-checked
ralphex:TEAM-03
Default Claude invocation bypasses permission prompts; safer host isolation is optional.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
ralphex docs/custom-providers.md
Finding: Claude Code is default; native --codex mode needs Codex 0.130.0+. Compatible wrapper scripts support Gemini, Pi, Copilot, OpenCode or custom tools. Native reviews use provider subagent mechanisms; --codex skips external Codex review.
-
Finding: Optional --worktree runs plans in separate Git worktrees; Docker wrapper mounts project/credentials with documented sandbox limits. Worktrees only apply to full/tasks-only modes.
ralphex:TEAM-05: The Ralphex dashboard displays task progress, logs and aggregate change counts; it does not render proposed-code diffs for human review.
ralphex:TEAM-05
The Ralphex dashboard displays task progress, logs and aggregate change counts; it does not render proposed-code diffs for human review.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: pending_reviews, evidence_navigation, comments, request_changes, shipping_guidance, effort, missed_evidence, keyboard, narrow_screen, context_switching, roles, retained_context.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:SHIP-01: Plan, commits, progress logs and reviewer output are retained. cross-checked
ralphex:SHIP-01
Plan, commits, progress logs and reviewer output are retained.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Optional browser dashboard streams SSE output with phase filters, collapsible sections, text search and late-join full history. Watch mode monitors multiple progress files in a session sidebar. This is monitoring, not established browser control of execution.
ralphex:SHIP-02: Default pipeline requests validation and multiple review phases; tasks-only and optional stages limit enforcement. cross-checked
ralphex:SHIP-02
Default pipeline requests validation and multiple review phases; tasks-only and optional stages limit enforcement.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:SHIP-03: Ralphex creates branches/worktrees and commits, and its optional finalize prompt rebases/organizes commits. It has no native PR creation or management operation.
ralphex:SHIP-03
Ralphex creates branches/worktrees and commits, and its optional finalize prompt rebases/organizes commits. It has no native PR creation or management operation.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: comments, ci_response, conflicts, product_state.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:SHIP-04: Finalize is optional customizable agent automation, not a tested delivery authority boundary. cross-checked
ralphex:SHIP-04
Finalize is optional customizable agent automation, not a tested delivery authority boundary.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:SHIP-05: Completed plans and commits persist; success has a documented archive caveat. cross-checked
ralphex:SHIP-05
Completed plans and commits persist; success has a documented archive caveat.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:OWN-01: Repository source is MIT licensed. cross-checked
ralphex:OWN-01
Repository source is MIT licensed.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: MIT license covers ralphex source at pinned revision.
ralphex:OWN-02: Core standalone CLI requires no product cloud service. cross-checked
ralphex:OWN-02
Core standalone CLI requires no product cloud service.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
ralphex:OWN-03: The inspected Ralphex configuration/runtime has no ADE analytics reporter or documented analytics opt-out control. Agent-provider traffic and optional user-configured completion notifications are separate.
ralphex:OWN-03
The inspected Ralphex configuration/runtime has no ADE analytics reporter or documented analytics opt-out control. Agent-provider traffic and optional user-configured completion notifications are separate.
- Documentation and pinned source inspection only; no hands-on scenario was run.
- Other atomic questions remain unestablished: code_data, prompts, credentials, logs, artifacts, retention, deletion, training, residency, subprocessors, telemetry_defaults, consent_prompt.
-
Finding: The inspected client/runtime route and command surface has terminal-started execution plus a read-only SSE monitoring dashboard, with no inbound scheduled/event execution, human prompt submission via shared sessions, human role/SSO layer or analytics opt-out switch. Dashboard diffStats contains only changed-file/addition/deletion counts, not proposed-code hunks. Review prompts expand reviewer roles into the selected harness native Task/spawn_agent tools; no Ralphex-owned session delegation tool is supplied. The optional finalize prompt rebases/organizes commits without a PR operation.
ralphex:OWN-04: ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage. cross-checked
ralphex:OWN-04
ralphex is MIT open source with no paid plan; it runs on your own Claude subscription (Agent SDK credit pool), Anthropic API key, or OpenAI/codex plan, and those providers bill for usage.
- First-party pricing pages, terms and licences read 2026-11-10; no purchase, billing or authenticated account test.
-
Finding: Says ralphex is open source under MIT and uses your own Claude or OpenAI plan. The site shows no price.
-
Finding: Says unattended runs count against your own Anthropic plan's Agent SDK credit pool or API-key billing, or can switch to your OpenAI/codex plan. ralphex itself has no paid tier and is MIT licensed.
-
Finding: The MIT License lets anyone obtain, use, copy and modify the software free of charge.
ralphex:OWN-06: Public release and issues exist, with explicit best-effort Windows support. cross-checked
ralphex:OWN-06
Public release and issues exist, with explicit best-effort Windows support.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.
-
Finding: Release published 7 September 2026; default-branch changes continue through 7 November.
ralphex:OWN-08: Open-source binary and source allow evaluation with user-provided agents/compute. cross-checked
ralphex:OWN-08
Open-source binary and source allow evaluation with user-provided agents/compute.
- Documentation establishes only the named facts. Unknown atomic questions and unrun scenarios remain open.
-
Finding: Standalone local CLI executes plan checklists, commits each task and runs configurable review stages. Linux/macOS supported; Windows source build is best-effort with missing signal/locking/process-cleanup behavior.